Anticoagulant therapies for the prevention of intravascular catheters malfunction in patients undergoing haemodialysis: systematic review and meta-analysis of randomized, controlled trials

Background. Catheter malfunction (CM), including thrombosis, is associated with reduced dialysis adequacy, as well as an increased risk of catheter-related bacteraemia (CRB) and mortality. The role of alternative anticoagulant regimens for CM prevention remains uncertain. Methods. A systematic review and meta-analysis were performed examining all randomized controlled trials (RCTs) assessing interventions acting via an anticoagulant mechanism compared with conventional care for the prevention of CM in adult patients receiving haemodialysis for end-stage kidney disease. Medline, EMBASE and the Cochrane Register were searched to November 2012. The primary outcome was CM. Secondary outcomes were CRB, all-cause mortality and bleeding events (all bleeding events reported or as defined by authors). Relative risks with 95% confidence intervals (CIs) for individual trials were pooled using random effects models for treatment classes. Results. The search yielded 28 trials including 3081 patients. Therapies assessed were alternative anticoagulant locking solutions (ALS), systemic warfarin and low/no dose heparin locking solutions (normal saline locks). No significant effect on CM (18 trials, 1579 participants) was observed for alternative ALSs (9 trials, 887 participants, RR 1.04, 95% CI 0.76-1.31), or low/no dose heparin (4 trials, 231 participants, RR 1.07, 95% CI 0.66-1.75), compared with heparin locking solutions (5000 units). Similarly, no significant effect was observed for warfarin (5 trials, 479 participants, RR 0.59, 95% CI 0.28-1.23) compared with placebo. A significant effect for CRB (15 trials, 2367 participants) was observed for ALSs (11 trials, 2010 participants, RR 0.41, 95% CI 0.30-0.56) but not for warfarin (1 trial, 174 participants, RR 2.44, 95% CI 0.89-6.55) or low/no dose heparin (3 trials, 183 participants, RR 1.01, 95% CI 0.47-2.19). All-cause mortality was not affected by ALSs (9 trials, 1719 participants, RR 0.83, 95% CI 0.56-1.23) or warfarin (3 trials, 403 participants, RR 0.78, 95% CI 0.37-1.65). Bleeding events were only reported in 7 trials, including only 2 trials of warfarin, with no clear effect demonstrated. Within the ALS group, the only agent with a reduction in CM was recombinant tissue plasminogen activator (rt-PA) locking solution (RR 0.58, 95% CI 0.37-0.91) based on the results of a single trial. Trials were mainly of high risk of bias. Conclusions. There is uncertainty on the benefits and harms of anticoagulant therapies over conventional care for prevention of CM. Further high-quality randomized trials, including safety outcomes, are needed.